The Next Generation of Stable Diffusion is upon us, and with it, I have managed to mix the perfect little mix to grant your creative visions the artistic freedom it deserves. Odem XL specializes in stylish, post-modern aestetics of tech wear, the shiny armors of fantasy, as well as just creating gorgeous characters. Being my first XL merge, expect some good updates to it in the not-so-far future, as I am still learning the ropes of XL merging. I hope you will enjoy creating with this as much as I did!<h1 id="heading-164">VERSION 2 AND NEWER ONLY:</h1>For higher quality outcomes, add the following text to your prompt:<pre><code>(very aesthetic, best quality, ultra detailed), intricate details</code></pre>Odem v2 uses a 'prompt reader' that actually understands the detail tags above. Thusly, it's now actually helpful to include these! Yay!Recommended Negative Prompt: (RESULTS WILL VARY DRASTICALLY IF YOU USE A DIFFERENT NEGATIVE PROMPT!!! LESS IS MORE!)low quality, (worst quality, normal quality:1.3)Recommended Settings:-DPM++ 2M Karras, DPM++ 3M SDE Karras and DMP++ 2M Exponential are my favourite Samplers for this model.-I use 35-45 steps on average for optimal results.-CFG from 8-12 recommended. Starting at CFG 13 eyes can start looking weird.-I prefer 4x_BooruGan_650k, 0,45 denoise and 20 hires steps. Nearest, Anime6B are also good options.-I prefer Anime6B upscaler for SD Upscale.-All of the showcase images have been generated in ComfyUI with 1 denoise, then upscaled with Boorugan 650k, and had a 2nd pass with 0,45-0,55 denoise with 20-25 steps and the same sampler settings.-Don't use Restore Faces.<h3 id="heading-717">License</h3>Odem XL from v2 onwards now uses the <a target="_blank" rel="ugc" href="https://freedevproject.org/faipl-1.0-sd/">Fair AI Public License 1.0-SD</a> inherited from Raemu XL 2.0, which inherited it from Animagine XL 3.0 compatible with Stable Diffusion models. Key points:1. Modification Sharing: If you modify Odem XL, you must share both your changes and the original license.2. Source Code Accessibility: If your modified version is network-accessible, provide a way (like a download link) for others to get the source code. This applies to derived models too.3. Distribution Terms: Any distribution must be under this license or another with similar rules.4. Compliance: Non-compliance must be fixed within 30 days to avoid license termination, emphasizing transparency and adherence to open-source values.<h3 id="heading-6">Note :</h3>To respect open source development, if you merge my model or Animagine XL 3.0, you need to provide details about the recipe parameters or training configuration if you are fine-tuning it, along with a statement about the license used. Thank you.Recipe:v2 - OdemV1 + Raemu XL with: 0.5,0.1,0.1,0.5,0.5,0.5,0.5,0.5,0.1,0.1,0.85,0.1,0.1,0.5,0.5,0.5,0.5,0.5,0.1,0.1then 0.5,0.1,0.1,0,0.1,0.5,0.5,0.5,0.1,0.1,0.85,0.1,0.1,0.5,0.5,0.5,0.1,0,0.1,0.1then merged both models with:0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5I used supermerger with merge block weights

Odem XL

odemXL_v10.safetensors

v1.0

v3.0

v2.1

Odem XL specializes in stylish, post-modern aestetics of tech wear, the shiny armors of fantasy, as well as just creating gorgeous characters. Being my first XL merge, expect some good updates to it in the not-so-far future, as I am still learning the ropes of XL merging. I hope you will enjoy creating with this as much as I did!Recommended Negative Prompt: (RESULTS WILL VARY DRASTICALLY IF YOU USE A DIFFERENT NEGATIVE PROMPT!!! LESS IS MORE!)low quality, (worst quality, normal quality:1.3)Recommended Settings:-DPM++ 2M Karras, Restart and DPM++ 2M Exponential are my favourite Samplers for this model.-I use 35-45 steps on average for optimal results.-CFG from 6-12 recommended. Starting at CFG 13 eyes can start looking weird.-I prefer NMKD-Siax, 0,45 denoise and 20 hires steps. BooruGan, Anime6B are also good options.-I prefer NMKD-Siax upscaler for SD Upscale.-All of the showcase images have been generated in ComfyUI with 1 denoise, then upscaled with Boorugan 650k, and had a 2nd pass with 0,45-0,55 denoise with 20-25 steps and the same sampler settings.-Don't use Restore Faces.

Files with SHA256: 8b980201a47ca556e30029aa72d3269fe20280c868015f7985e685071070e4df